We include the 'label.txt' of the randomly selected examples.
The text before '|' in each line is the original transcription and the text after '|' is the targeted transcription.